Letter
pubs.acs.org/NanoLett
Surface-Enhanced Raman Scattering Plasmonic Enhancement Using
DNA Origami-Based Complex Metallic Nanostructures
M. Pilo-Pais,*,† A. Watson,† S. Demers,† T. H. LaBean,‡ and G. Finkelstein*,†
†
Department of Physics, Duke University, Durham, North Carolina 27708, United States
Department of Materials Science and Engineering, North Carolina State University, Raleigh, North Carolina 27607, United States
‡
S Supporting Information
*
ABSTRACT: DNA origami is a novel self-assembly technique allowing one to form various two-dimensional shapes and
position matter with nanometer accuracy. We use DNA origami templates to engineer surface-enhanced Raman scattering
substrates. Specifically, gold nanoparticles were selectively placed on the corners of rectangular origami and subsequently
enlarged via solution-based metal deposition. The resulting assemblies exhibit “hot spots” of enhanced electromagnetic field
between the nanoparticles. We observed a significant Raman signal enhancement from molecules covalently attached to the
assemblies, as compared to control nanoparticle samples that lack interparticle hot spots. Furthermore, Raman molecules are
used to map out the hot spots’ distribution, as they are burned when experiencing a threshold electric field. Our method opens
up the prospects of using DNA origami to rationally engineer and assemble plasmonic structures for molecular spectroscopy.
KEYWORDS: DNA origami, Raman, SERS, 4-aminobenzenethiol
utilizing the regions of intense electric field created near
granular metallic surfaces. These “hot spots” can be understood
as resulting from localized surface plasmon modes resonantly
excited by the incident laser. The analyte molecules that happen
to be positioned in the hot spots provide disproportionately
high contribution to the Raman scattering, resulting in a signal
enhancement that is many orders of magnitude.19,20
Pairwise complementary DNA strands can be used to
position Raman-active molecules between functionalized gold
nanoparticles (AuNPs), thus fabricating plasmonic structures
with active hot spots.21−24 In this paper, we utilized DNA
origami to fabricate more complex multiparticle assemblies, and
determined their performance as SERS substrates.
Specifically, we used DNA origami to organize the metallic
structures and then covalently attached Raman-active molecules
to the metal. We found that the substrates with four
nanoparticles (NPs) per origami produce a strongly enhanced
Raman signal compared to the control samples with only one
D
NA origami is a product of a one-pot reaction in which
DNA strands of specific sequences self-assemble into a
large structure (∼100 nm) of a predetermined shape,1 thereby
providing an alternative to conventional top-down fabrication
methods. The resulting templates are highly addressable and
versatile tools for site-specific placement of various nanocomponents, such as metallic nanoparticles,2−6 quantum dots,7
fluorophores,8 and carbon nanotubes.9 It has also been shown
that origami templates can serve as platforms for DNA
motors,10 DNA walkers,11−13 and chemical reactions.14 Most
recently, origami templates have been used to promote
interactions between attached nanocomponents, such as
plasmonic coupling between gold nanorods4 or gold nanoparticles,15 as well as enhancement and quenching of
fluorophores16,17 or CdSe quantum dots (QDs).18
Building on the massively parallel formation of DNA origami
and their capability to serve as a nanobreadboard, one can
further envision using them as biosensors. One particularly
attractive goal is to facilitate Raman spectroscopy, which
provides a highly specific chemical fingerprint. Unfortunately,
the Raman scattering cross section is small; surface-enhanced
Raman spectroscopy (SERS) greatly increases the signal by
© 2014 American Chemical Society
Received: January 24, 2014
Revised: March 1, 2014
Published: March 19, 2014
2099
dx.doi.org/10.1021/nl5003069 | Nano Lett. 2014, 14, 2099−2104
Nano Letters
Letter
Figure 1. Typical SERS spectra of 4-aminobenzenethiol (4-ABT) attached to metal nanoparticles assembled on DNA origami. (A) Four-particle
assemblies (“tetramers”) which have interparticle hot spots; (B) control sample with one AuNP per origami (“monomers”) that lack the interparticle
hot spots. Spectra correspond to the first 1 s of the laser exposure and are normalized to the average density of nanoparticles in the illuminated
region. Insets: SEM images taken from the measurement areas and cartoons representing the target structure. The origami template is shown in blue,
while the red tint in (A) indicates the regions of enhanced electric field (hot spots). Larger area images and higher-magnification views of the
individual structures can be found in the Supporting Information.
used to assemble complex metallic structures for optical
applications such as SERS substrates.
The structures were enlarged using a commercial silver
metallization kit following the manufacturer directions (HQ
silver enhancement, Nanoprobes INC), as performed in our
previous works.6,28 The optimal incubation time was 12 min for
the tetramer structures. This was the necessary time to have
most of the nanoparticles closely spaced but typically not
touching, permitting the formation of interparticle hot spots.
The control samples were incubated for 10 min to achieve the
same average NP size of 50 nm. Although measuring the gap
between nanoparticles with nanometer precision is challenging,
we estimate the average gap size to be below 3 nm (Supporting
Information Figure S2). This separation distance was clearly
small enough to produce a strong signal enhancement, as
demonstrated from the Raman spectra (Figure 1). We
lithographically patterned the samples with a set of markers,
which allowed us to identify specific regions where the spectra
were taken. Specifically, we spin-coated PMMA-A4 (MicroChem Corp.), followed by a 5 min UV exposure and
development to open ∼10 μm windows. The samples were
then imaged using a scanning electron microscope (SEM); we
selected for further study the windows that showed no
abnormally assembled structures such as the multiparticle
clusters seen in Supporting Information Figure S2. We also
measured the average nanoparticle density in each window,
used for normalization purposes. The Raman spectra exhibit no
discernible difference before or after lithographic patterning.
The samples were then incubated in a 5 mM solution of 4aminobenzenethiol (4-ABT) in ethanol for two hours (long
enough to reach full surface coverage). Placement of multiple
Raman molecules throughout the hot spot is imperative to
obtain an estimate of the enhancement factor, which is not
possible with a single molecule. The thiol functional group
ensures that the Raman-active molecules are covalently
attached only to the metallic surfaces. The samples were then
thoroughly rinsed in pure ethanol to remove any physisorbed
molecules. Similarly treated SiO2 substrates without NPs
showed no detectable 4-ABT Raman signal, indicating the
effectiveness of the rinsing procedure. Any potential carbon
contamination due to SEM imaging is removed during 4-ABT
nanoparticle per origami. Indeed, the small gaps between
closely spaced nanoparticles result in hot spots, which are
absent in samples with individual nanoparticles (Figure 1).
Furthermore, the Raman signal systematically decayed as a
function of the laser exposure time in the samples with four
particles per origami. We attribute this behavior to molecular
damage caused by the high electric field at the hot spots. The
one-particle control samples lacking the interparticle hot spots
exhibited no such decay.
Results and Discussion. We use DNA origami to control
the composition, shape, geometry, and arrangement of metallic
structures, which in turn determine the local distribution of
electromagnetic fields. DNA-metallic assemblies were prepared
as we previously reported.6 Briefly, select staple strands of the
standard “tall” rectangular DNA origami (∼90 × 70 nm2) are
extended by a specific ss-DNA sequence, referred to as X24. The
sequence serves as an anchor for AuNPs; to increase the
binding probability the anchors were positioned in pairs on
adjacent staples. The AuNPs are conjugated with ∼5
complementary sequence strands (X24,comp) through standard
thiol chemistry25 (please refer to Methods for complete
experimental details and to the Supporting Information for
the list of sequences used on the modified DNA strands).
Two different types of SERS samples were prepared. In the
sample with engineered hot spots, four AuNPs were attached to
each of the four corners of the origami template (tetramer
samples); in the control samples, only one AuNP was placed in
one of the corners (monomer samples). Each of the modified
DNA template designs was attached to RCA cleaned (SC-1 and
SC-2) and oxygen plasma ashed (SPI Plasma Prep II, 20 min,
100 mA) silicon dioxide substrates (1 μm oxide, Silicon Quest)
using a 10× TAE, 125 mM Mg2+ solution (the final DNA
origami concentration was 250 pM). Functionalized AuNPs
were then added to the solution (final concentration of 3 nM)
resulting on the attachment to the origami templates. The
samples were then incubated for 15 min and rinsed in DI water
for 15 s, followed by gentle drying with nitrogen. We chose to
assemble tetramers as opposed to dimers because they exhibit
interesting plasmonic properties such as Fano resonances,26
which have been shown to provide much greater SERS
enhancement due to near-field intensity variations.27 The
present work is a proof of concept that DNA origami can be
2100
dx.doi.org/10.1021/nl5003069 | Nano Lett. 2014, 14, 2099−2104
Nano Letters
Letter
Figure 2. (A) Raman spectra taken by repeatedly exposing the tetramer sample to 1 s HeNe laser pulses. (B) Intensity of the 1075 and 1141 cm−1
Raman peaks (background subtracted) and the fluorescence background as a function of the laser exposure time. The rapid decay is attributed to the
photodamage of the molecules caused by the intense field at the hot spots.
Figure 3. (A) Raman spectra in the vicinity of the 950 cm−1 substrate band and the 1075 cm−1 4-ABT peak. Each spectrum is taken after increasing
the laser intensity to 1, 10, 25, 50, and 100% of I0 and waiting for the signal to saturate at the new intensity level. Two or three spectra are
consecutively taken at each intensity level, demonstrating the saturation of the signal. (B) Ratio of the time-saturated 4-ABT Raman peak
(background subtracted) to the laser intensity I as a function of (I0/I)1/2 ∝ 1/E. The colored data sets correspond to different locations on the
sample and are scaled to 1 at I = I0. (C) Cartoon showing the electric field enhancement expansion of the regions within the hot spot, as a function
of the electric field intensity. Optical filters are interchanged in order to increase the incident field in steps, burning the molecules located in
progressively larger regions of the hot spots.
ethanolic incubation.29 Incidentally, we did not observe Raman
signatures of any other molecules, such as DNA.
The Raman spectra were obtained using a Jobin Yvon
LabRam ARAMIS (Horiba, Ltd.) spectrometer using a 632.8
nm, 5 mW HeNe laser excitation, focused by a 100× objective
to a ∼1 μm spot. Figure 1 compares the Raman signal
measured from tetramer and monomer samples during the first
second exposure at maximal laser intensity (I0). The spectra
were normalized to the average NP number in the
corresponding lithographic window. We verified that the
magnitude of Raman signal per particle was reproducible in
other regions. The average relative Raman enhancement per
NP is ∼100 in tetramer versus monomer samples. Note that
using the monomer samples as a control ensures that the
surface concentration of the covalently attached 4-ABT layer is
the same as in the tetramer samples. This eliminates the
uncertainty in determining the SERS enhancement factor,
common for the measurements that use molecular solutions as
a control.19
The observed SERS enhancement is naturally explained by
the hot spots created in the tetramer sample. Numerical
simulations (not shown) indicate that the hot spots are located
between pairs of NPs (see cartoon in Figure 3C) similar to the
hot spots created in nanoparticle dimers. Unlike dimers, where
the enhancement disappears for an electric field perpendicular
to the dimer axis, the hot spots in the tetramers should be
2101
dx.doi.org/10.1021/nl5003069 | Nano Lett. 2014, 14, 2099−2104
Nano Letters
Letter
relative prominence of the 950 cm−1 substrate band and allows
us to visualize the effect).
In Figure 3B, we plot the ratio of the 4-ABT Raman signal to
the laser intensity, I. The horizontal axis is (I0/I)1/2,
proportional to the inverse incident electric field.33,34 Data
taken at three different spots are represented by different
colors, all normalized to 1 at I = I0. In each of the curves, the
relative Raman signal at I = 0.01I0 is 20−40 times stronger than
the signal at I0. In other words, at the maximal laser intensity
the high fields of the hot spots damage the molecules
responsible for most of the Raman enhancement that was
achieved at 1% of the intensity. Still, we recall that at I0 the
signal measured from the tetramer sample is ∼100 times
stronger compared to the monomers. Because the threshold
photodamage field is not yet reached in the monomer samples,
even at I0 their Raman signal should scale proportionally to the
laser intensity. Therefore, at 1% of I0 the enhancement factor of
the tetramer could potentially reach 2000−4000. Unfortunately, the direct comparison was not feasible, because the
Raman signal of the monomer sample was too weak at 0.01I0,
while at I0 the Raman signal from the tetramer samples
experiences a very significant degradation prior to the
completion of the first 1 s exposure.
Conclusion. We have shown that DNA origami can be
successfully used to engineer substrates for surface-enhanced
Raman spectroscopy. The Raman signal of 4-ABT molecules
deposited on the tetramer NP assemblies is enhanced at least a
hundred times (and potentially several thousand) per nanoparticle as compared to control samples with individual
nanoparticles. The enhancement is due to hot spots, whose
existence was verified by time and intensity-dependent
measurements. Our results demonstrate the design capabilities
that origami-based metallic structures can offer for spectroscopic and plasmonic applications. In the future, we plan to
fully use the addressability of the DNA origami and to custom
tune the plasmon resonance frequency of the DNA-metallic
structures by exploring different shapes and materials. This will
allow us to tailor the plasmonic resonances to match the laser
frequency and/or the optimal excitation band of a given
molecule. The methodology presented here opens up new
possibilities to rationally engineer substrates using DNA
origami for optical and plasmonic applications.
Methods. DNA Templates Synthesis. All DNA sequences
were purchased from IDT (Integrated DNA Technologies,
Inc.). The modified tall rectangular DNA origami templates
were formed using Rothemund design1 but with the following
modifications: All of the side staples were left out to prevent
stacking between multiple origamis. The binding sites the
AuNPs anchor were made by extending two consecutive staples
with a short spacer sequence T5 followed by a 24 bp DNA
sequence (X24) (see Supporting Information for a list of DNA
sequences). The prepared origami (∼5 nM) was filtered from
the excess staples by centrifuging using a ultrafiltration
centrifuge unit (100 KDa MWCO, Millipore) with three
washes of 1× TAE, 12.5 mM Mg2+ buffer.
AuNP DNA Conjugation. Gold nanoparticles were concentrated and functionalized using a phosphine recipe originally
developed by ref 25 but with some changes: 10 mL of 5 nm, 80
nM AuNPs solution (British Biocell International) were
incubated overnight with 3 mg of bis(p-sulfonatophenyl)phenylphosphine (BSPP, Sigma-Aldrich). The AuNPs were
concentrated by adding 250 mg of NaCl and centrifuging for 30
min at 800 g. The supernatant was removed without disturbing
activated by any laser polarization. Our control monomer
samples lack the interparticle hot spots; although the electric
field is also enhanced at the particle poles, the enhancement
factor should be much smaller, and we disregard it in the
following discussion. Notice also that the 1141 cm−1 Raman
mode (blue shadow in Figure 1) is no longer observed from the
monomer samples. This can be attributed to a strong chemical
dependence of the Raman mode, which requires a minimum
excitation energy in order to promote a charge transfer.30,31
The importance of the hot spots is further evidenced by the
time evolution of the Raman signal. Figure 2A shows the
sequence of spectra taken during successive 1 s exposures. The
signal intensity initially drops rapidly and then saturates. Similar
intensity decay has been attributed to photodamaging of
Raman molecules by the enhanced field at the hot spots.32−34
Alternatively, the decay has been assigned to morphological
changes of the metallic structures due to heating.35 Although it
is difficult to identify the mechanism responsible for the decay
we observe,19 we tentatively attribute it to molecular photodamage. SEM images of the structures before and after the
Raman measurement do not show noticeable differences,
within a resolution of a few nanometers. It is also important to
emphasize that the molecules are covalently attached to the
silver particles through the thiol cross-linker, preventing them
from leaving the hot spots (photodesorption).
Figure 2B shows the integrated Raman signal for the 1075
cm−1 (CS stretch, 7a1) and the 1141 cm−1 (CH bend, b2)
peaks, as well as the fluorescence background as a function of
the laser exposure time. All the signals are normalized to the
values measured at the first 1 s exposure. The signal decays
rapidly at first and then saturates at a constant value. The inset
of Figure 2B shows the integrated intensity of the 1075 cm−1
peak measured from the control monomer sample; no signal
decay is observed in this case due to the lack of interparticle hot
spots.
The saturation of the Raman signal after about 100 s allows
us to further characterize the electric field enhancement in the
hot spots.32,33 We assume that only the molecules experiencing
an electric field exceeding a certain critical value are
destroyed.19 No signal decay is observed at 1% of the maximal
laser intensity I0, indicating that the critical field is not yet
reached even in the hot spots. To study the successive
photodamage of the hot spots by the laser, we increased the
illumination intensity to 10, 25, 50 and 100% of I0. Each
successive increase of the intensity is followed by the gradual
decay of the Raman signal, similar to the decay shown in Figure
2. This behavior indicates the stepwise expansion of the regions
where the field exceeds the critical value and the molecules are
photodamaged (see schematics in Figure 3C). As a result, the
ratio of the time-saturated Raman signal to the laser intensity
decreases with increasing intensity.
Figure 3A illustrates this behavior by showing the timesaturated spectra for each subsequent intensity increment, all
measured from the same spot. Notice the decay of the relative
intensity of the 4-ABT Raman signal at 1075 cm−1 as compared
to the substrate Raman band centered at 950 cm−1. The latter
signal was verified to scale proportionally to the laser intensity,
while the 4-ABT signal is clearly sublinear; for example, at the
full intensity of I0 the 4-ABT Raman peak is barely visible
relative to the 950 cm−1 band, while at 1% of I0 the 4-ABT
Raman peak dominates (this figure is not normalized to the
number of particles. Also, it is measured from a more dilute
sample compared to Figures 1 and 2, which increases the
2102
dx.doi.org/10.1021/nl5003069 | Nano Lett. 2014, 14, 2099−2104
Nano Letters
■
the AuNP pellet and resuspended on 200 μL of methanol and
200 μL of BSPP solution (3 mg in 10 mL DI water). The
solution was once again centrifuged for 30 min and its
supernatant removed. The AuNPs were resuspended with 200
μL of the same BSPP solution and incubated for 48 h with
disulfide-modified X24 DNA sequence at a ratio of 1:5 Au:DNA
and adjusted to 1× TAE, 50 mM NaCl. Thiolated T5 strands
were added at a Au/T5 ratio of 1:60 to fully backfill the AuNPDNA conjugates in order to prevent aggregation in a 125 mM
Mg2+ environment. Excess DNA strands were removed by
running the AuNP-DNA conjugates on a 3% agarose gel (0.5×
TAE) for 25 min at 10 V/cm. As indicated in ref 6, we find this
purification step critical to obtain a high-yield binding. The
AuNP-conjugate was recovered using Freeze and Squeeze (BioRad Laboratories) with a typical recovery concentration of 500
nM.
DNA-Metallic Structures and Individual Nanoparticles
Formation. The modified DNA templates were attached to
previously RCA cleaned and oxygen plasma ashed (SPI Plasma
Prep II, 20 min, 100 mA) silicon dioxide substrates (1 μm
oxide, Silicon Quest) using 10× TAE, 125 mM Mg2+ solution.
The final DNA origami concentration was 250 pM. Functionalized gold nanoparticles (AuNPs) were attached onto the
templates by adding 1 μL of a concentrated solution, to a final
concentration of 3 nM and incubated for 15 min and rinsed in
water for 5 s, followed by nitrogen blow. The structures were
then enlarged in size using in-solution silver metallization as
indicated by the manufacturer (HQ silver enhancement,
Nanoprobes INC) for 12 min for the DNA-metallic structures
and 10 min for the individual nanoparticles.
Lithography. Samples were spin coated with PMMA-A4
(MicroChem Corp) and baked on a hot plate (120 C surface
temperature) for 2 min. A copper grid of 2000 mesh (Structure
Probe, Inc.) was placed on top and the sample was exposed to
UV (500 W) for 5 min.
Raman Measurements. DNA-metallic structures were
incubated on a 5 mM 4-aminobenzenethiol ethanolic solution
for 4 h followed by a thorough rinse on pure ethanol. Raman
spectra were obtained using a Jobin Yvon LabRam ARAMIS
(Horiba, Ltd.) Raman instrumentation. Measurements were
taking using a 5 mW HeNe 632.8 nm laser using 1 s intervals. A
100× objective, resolution grating of 1800 grooves, and a slit of
100 μm were used on all measurements. The spectra ranged
from 1000 to 1600 cm−1.
■
ACKNOWLEDGMENTS
The authors thank Henry Everitt for his useful suggestions. M.
Pilo-Pais thanks Jack Mock and Christos Argyropoulos for
valuable discussions on optical measurements. This work has
been supported by NSF-ECCS-12-32239.
■
REFERENCES
(1) Rothemund, P. W. K. Nature 2006, 440, 297−302.
(2) Ding, B.; Deng, Z.; Yan, H.; Cabrini, S.; Zuckermann, R. N.;
Bokor, J. J. Am. Chem. Soc. 2010, 132, 3248−3249.
(3) Hung, A. M.; Micheel, C. M.; Bozano, L. D.; Osterbur, L. W.;
Wallraff, G. M.; Cha, J. N. Nat. Nanotechnol. 2010, 5, 121−126.
(4) Pal, S.; Deng, Z.; Ding, B.; Yan, H.; Liu, Y. Angew. Chem., Int. Ed.
2010, 49, 2700−2704.
(5) Pal, S.; Varghese, R.; Deng, Z.; Zhao, Z.; Kumar, A.; Yan, H.; Liu,
Y. Angew. Chem., Int. Ed. 2011, 123, 4262−4265.
(6) Pilo-Pais, M.; Goldberg, S.; Samano, E.; LaBean, T. H.;
Finkelstein, G. Nano Lett. 2011, 11, 3489−3492.
(7) Bui, H.; Onodera, C.; Kidwell, C.; Tan, Y.; Graugnard, E.; Kuang,
W.; Lee, J.; Knowlton, W. B.; Yurke, B.; Hughes, W. L. Nano Lett.
2010, 10, 3367−3372.
(8) Steinhauer, C.; Jungmann, R.; Sobey, T.; Simmel, F.; Tinnefeld,
P. Angew. Chem., Int. Ed. 2009, 48, 8870−8873.
(9) Maune, H. T.; Han, S.-p.; Barish, R. D.; Bockrath, M.; Goddard,
W. A., III; Rothemund, P. W. K.; Winfree, E. Nat. Nanotechnol. 2010,
5, 61−66.
(10) Venkataraman, S.; Dirks, R. M.; Rothemund, P. W. K.; Winfree,
E.; Pierce, N. A. Nat. Nanotechnol. 2007, 2, 490−494.
(11) Gu, H.; Chao, J.; Xiao, S.-J.; Seeman, N. C. Nature 2010, 465,
202−205.
(12) Lund, K.; Manzo, A. J.; Dabby, N.; Michelotti, N.; JohnsonBuck, A.; Nangreave, J.; Taylor, S.; Pei, R.; Stojanovic, M. N.; Walter,
N. G.; Winfree, E.; Yan, H. Nature 2010, 465, 206−210.
(13) Wickham, S. F. J.; Endo, M.; Katsuda, Y.; Hidaka, K.; Bath, J.;
Sugiyama, H.; Turberfield, A. J. Nat. Nanotechnol. 2011, 6, 166−169.
(14) Voigt, N. V.; Torring, T.; Rotaru, A.; Jacobsen, M. F.;
Ravnsbaek, J. B.; Subramani, R.; Mamdouh, W.; Kjems, J.; Mokhir, A.;
Besenbacher, F.; Gothelf, K. V. Nat. Nanotechnol. 2010, 5, 200−203.
(15) Kuzyk, A.; Schreiber, R.; Fan, Z.; Pardatscher, G.; Roller, E.-M.;
Hogele, A.; Simmel, F. C.; Govorov, A. O.; Liedl, T. Nature 2012, 483,
311−314.
(16) Acuna, G. P.; Bucher, M.; Stein, I. H.; Steinhauer, C.; Kuzyk, A.;
Holzmeister, P.; Schreiber, R.; Moroz, A.; Stefani, F. D.; Liedl, T.;
Simmel, F. C.; Tinnefeld, P. ACS Nano 2012, 6, 3189−3195.
(17) Acuna, G. P.; MÃ ű ller, F. M.; Holzmeister, P.; Beater, S.;
Lalkens, B.; Tinnefeld, P. Science 2012, 338, 506−510.
(18) Ko, S. H.; Du, K.; Liddle, J. A. Angew. Chem., Int. Ed. 2013, 52,
1193−1197.
(19) Ru, E. L.; Etchegoin, P. Principles of Surface-Enhanced Raman
Spectroscopy, and related plasmonic effects, 1st ed.; Elsevier Science:
Oxford, 2009.
(20) Kleinman, S. L.; Frontiera, R. R.; Henry, A.-I.; Dieringer, J. A.;
Van Duyne, R. P. Phys. Chem. Chem. Phys. 2013, 15, 21−36.
(21) Cao, Y. C.; Jin, R.; Mirkin, C. A. Science 2002, 297, 1536−1540.
(22) Graham, D.; Thompson, D. G.; Smith, W. E.; Faulds, K. Nat.
Nanotechnol. 2008, 3, 548−51.
(23) Lim, D.-K.; Jeon, K.-S.; Kim, H. M.; Nam, J.-M.; Suh, Y. D. Nat.
Mater. 2010, 9, 60−67.
(24) Lim, D.-K.; Jeon, K.-S.; Hwang, J.-H.; Kim, H.; Kwon, S.; Suh, Y.
D.; Nam, J.-M. Nat. Nanotechnol. 2011, 6, 452−460.
(25) Loweth, C. J.; Caldwell, W. B.; Peng, X.; Alivisatos, A. P.;
Schultz, P. G. Angew. Chem., Int. Ed. 1999, 38, 1808−1812.
(26) Shafiei, F.; Monticone, F.; Le, K. Q.; Liu, X.-X.; Hartsfield, T.;
Alù, A.; Li, X. Nat. Nanotechnol. 2013, 8, 95−9.
(27) Ye, J.; Wen, F.; Sobhani, H.; Lassiter, J. B.; Van Dorpe, P.;
Nordlander, P.; Halas, N. J. Nano Lett. 2012, 12, 1660−7.
(28) Park, S. H.; Prior, M. W.; LaBean, T. H.; Finkelstein, G. Appl.
Phys. Lett. 2006, 89, 033901−3.
ASSOCIATED CONTENT
S Supporting Information
*
DNA sequences and additional SEM images. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
Letter
AUTHOR INFORMATION
Corresponding Authors
*E-mail: (M.P.P.) mgp8@phy.duke.edu.
*E-mail: (G.F) gleb@phy.duke.edu.
Author Contributions
M.P., T.H.L., and G.F. designed the experiment. M.P., A.W.,
and S.D. fabricated the samples and conducted the experiment.
M.P., A.W., and G.F. analyzed and interpreted the data.
Notes
The authors declare no competing financial interest.
2103
dx.doi.org/10.1021/nl5003069 | Nano Lett. 2014, 14, 2099−2104
Nano Letters
Letter
(29) Moran, C. H.; Xia, X.; Xia, Y. Phys. Chem. Chem. Phys. 2013, 15,
5400−6.
(30) Kim, K.; Lee, H. B.; Shin, D.; Ryoo, H.; Lee, J. W.; Shin, K. S. J.
Raman Spectrosc. 2011, 42, 2112−2118.
(31) Kim, K.; Kim, K. L.; Shin, D.; Choi, J.-Y.; Shin, K. S. J. Phys.
Chem. C 2012, 116, 4774−4779.
(32) Etchegoin, P. G.; Le Ru, E. C.; Maher, R. C.; Cohen, L. F. Phys.
Chem. Chem. Phys. 2007, 9, 4923−4929.
(33) Fang, Y.; Seong, N.-H.; Dlott, D. D. Science 2008, 321, 388−
392.
(34) Yang, Y.; Callahan, J. M.; Kim, T.-H.; Brown, A. S.; Everitt, H.
O. Nano Lett. 2013, 13, 2837−2841.
(35) Soudamini, D. U.; Fu, C. Y.; Kho, K. W.; Thoniyot, P.; Agarwal,
A.; Olivo, M. Proc. SPIE 2009, 7394, 73940−9.
2104
dx.doi.org/10.1021/nl5003069 | Nano Lett. 2014, 14, 2099−2104